While the feature set looks comprehensive, the details around cost control could be clearer for some users. The site mentions daily channel limits, token totals, and OAuth quota windows, but it does not explicitly spell out how these limits interact or what happens when you exceed them. For instance, does hitting a daily cost limit immediately stop requests, or is there a grace period? The distinction between standard cost tracking and sandbox fallbacks for Google Cloud Code channels is mentioned, but the exact financial implications are not fully detailed. Additionally, while the admin console exposes runtime state, it is not entirely clear if historical data beyond seven days is retained or if that is a hard limit for the free tier. For someone considering this company for long-term use, understanding the exact scope of these limits and the consequences of overages would be important before committing. The system seems powerful, but the pricing and limit structures require a closer look at the documentation to ensure they align with specific budget expectations.
Questions about ccLoad
What is ccLoad?
ccLoad is an open-source, self-hosted AI API gateway designed for Claude Code, Codex, and Gemini clients. It sits between clients and model providers to manage API keys and account credentials behind a single base URL. The platform handles failover before the first byte and meters costs on every request.
How does ccLoad handle upstream failures?
The system uses priority-based and weighted round-robin distribution to route requests. It automatically fails over when an upstream fails, isolating issues at the key, model, channel, or URL level. This ensures that a single upstream failure does not disrupt an entire coding session.
What integrations and protocol transformations are available?
Users can import credentials from providers like Anthropic and Codex using OAuth channel support. The gateway also supports protocol transformation and fallback between HTTP/SSE and WebSocket. This allows Codex Responses over WebSocket to be used alongside standard HTTP requests.
How does ccLoad manage costs and pricing?
Cost control features include daily channel limits and token usage limits tracked in micro-dollars. The pricing model includes cache, long-context tiers, and service_tier with daily and monthly caps per token. These measures help keep the bill predictable for long-running sessions.
What monitoring and error handling features exist?
The platform provides real-time monitoring of service health, client-protocol statistics, and cost analysis. It also offers failover before output, moving to the next channel on 429s, 5xx errors, or errors disguised as HTTP 200. This prevents the client from seeing incomplete or failed outputs.
How do users configure and manage channels?
Administrators can add channels in the admin console by pasting API keys or importing account credentials. They can issue access tokens for each client or teammate with optional model, channel, cost, and concurrency limits. Upstream keys never leave the gateway, enhancing security.
Which subscription accounts and models are supported?
The gateway supports subscription accounts from Codex, Claude, Antigravity, xAI, CodeBuddy, Z.ai, Cursor, and Zed. Tokens refresh automatically, and quota usage is tracked against each provider's reset window. Any client on any upstream can use every channel that serves the model.
What do reviewers praise about ccLoad?
Some reviewers praise the ability to failover before the first byte, ensuring sessions stay alive during long coding tasks. Others appreciate the cost metering that tracks usage in micro-dollars to prevent unexpected bills. The open-source nature and MIT license are also frequently mentioned positively.
Who is ccLoad best suited for?
The platform is specifically built for coding agents that stream long responses and run for hours. It is designed for users who need to keep sessions alive and maintain predictable billing. It is not a plain API relay but a specialized gateway for these demanding workflows.
How can a user get started with ccLoad?
To get started, users can deploy the open-source software on their own server. They connect upstream channels via the admin console and issue tokens for their clients. The project is available on GitHub under an MIT license for free use.
How does protocol conversion work?
The system converts the protocol when the upstream and client protocols differ. It handles requests and streamed responses even when the protocols are not identical. This flexibility allows Claude Code, Codex, and Gemini clients to work seamlessly together.
What are the key configuration options?
The gateway tracks usage against each provider's reset window rather than just API keys. It also supports long-context tiers and service_tier configurations. These features are essential for managing the unique demands of AI coding agents.
What are the core technical specifications?
The website states that ccLoad is open-source and self-hosted. It is compatible with OpenAI-compatible SDKs and tools. The project includes documentation, a Docker image, and a public source code repository.
What is the official website of ccLoad?
The official website of ccLoad is ccload.xyz.
How is ccLoad rated on Trustburn?
On Trustburn, ccLoad has a rating of 4 out of 5 based on 4 reviews.
AI-generated from reviews on this page. Verify anything important with the company directly.
It is really convenient that you can deploy via Docker, Hugging Face, or just download a binary. That flexibility makes testing and production setup much easier for different team sizes.
The availability of Docker images and Hugging Face hosting makes getting started very straightforward.
The automatic failover logic described on the site appears highly effective for maintaining service continuity. By isolating errors at the key or model level, ccLoad prevents a single issue from cascading across all channels. The real-time monitoring dashboard provides valuable data on cost analysis and client-protocol statistics, which is essential for managing API usage efficiently. Additionally, the support for various deployment options, including Docker and Hugging Face, suggests flexibility for different technical environments. The system seems designed to handle complexity well, particularly with its protocol transforms and smart routing capabilities. This makes it a strong candidate for teams managing multiple AI model integrations simultaneously.
News
No news yet. This is where the business shares its own updates and press.
Photos
Widgets
Get the widget for your site →About ccLoad
ccLoad is an AI API gateway designed for use with Claude Code, Codex, Gemini, and OpenAI-compatible clients. It provides smart routing with priority-based and weighted round-robin distribution, along with automatic failover that isolates failures at the key, model, channel, or URL level. The platform offers real-time monitoring of service health, client-protocol statistics, and cost analysis, as well as cost control features such as daily channel limits and token usage limits tracked in micro-dollars. Additional capabilities include OAuth channel support for importing credentials from providers like Anthropic and Codex, protocol transformation and fallback between HTTP/SSE and WebSocket, and per-channel proxy and scheduling options.
- Website
- ccload.xyz